Copyright © 2021 Cryptolaw. All Rights Reserve.
Concerns over market manipulation were cited by the US SEC as one of the reasons not to approve bitcoin ETFs thus far. Public officials around the world have voiced similar concerns over abusive market practices, such as frontrunning, wash trades and insider trading.
So what do we know so far? How big is this problem?
Back in 2018, the OAG NY published the report of its Virtual Markets Integrity Initiative. It focused on centralized exchanges and summarized concerns and responses received from exchanges that filled in the questionnaire. The OAG listed a number of concerns, such as conflicts of interest (employee and proprietary trading, listing fees) and market manipulation.
Wash trades or volume inflation
In March 2019, Bitwise Asset Management made a presentation to the US SEC (in the context of its Bitcoin ETF Trust application) in which it claimed that roughly 95% of alleged trading volume on bitcoin exchanges was fake or wash trading. It pointed fingers at the widely used bitcoin trading volume data from CoinMarketCap.com, calling that data simply ‘wrong’.
Later that year, CoinMarketCap introduced changes to the metrics for its ratings and it has tweaked those metrics several times since.
In a recent draft paper, researcher analysed trading data between July-Nov. 2019 on 29 crypto exchanges. For unregulated exchanges (as defined by the authors), they found more than 70% of reported trading volume (in bitcoin, ether, litecoin and XRP) were wash trades, worth an estimated $4.5 trillion in spot markets (and $1.5 trillion in derivatives markets) in Q1 2020.
Wash trading/volume inflation practices were linked to centralized exchanges (rather than decentralized exchanges or DEXes). DeFi’ers would probably point out that this shows why we need decentralized services, with greater on-chain transparency.
However, DeFi (decentralized finance) hasn’t been spared of allegations of abusive market practices. More precisely: there have been lots of discussions about frontrunning on the Ethereum blockchain, on which most of DeFi is currently built.
A widely discussed paper, Flashboys 2.0 by computers scientist including Philip Daian and Ari Juels, laid out the prevalence of frontrunning in the Ethereum mempool (the pool of unconfirmed transactions).
In a nutshell: users submit transactions to the Ethereum network and pay a proposed ‘gas fee’ (transaction fee) to incentivize miners to include the transaction in a new block on the Ethereum blockchain. The transaction is first sent to the mempool (the pool of unconfirmed transactions on Ethereum), waiting for a miner to pick it up and include it into a new block.
Miners are free to choose which unconfirmed transactions in the mempool they wish to select to include in a proposed new block. Miners typically prioritize transactions that include a higher ‘gas fee’ (transaction fee). This is the most lucrative strategy.
There is a time lag between the submission of a transaction to the Ethereum mempool and the time the transaction is included in a block that is appended to the Ethereum blockchain. During that time, unconfirmed transactions are just laying idle, hanging around in the hope of being mined. That time lag opens the door to frontrunning. While unconfirmed transactions are laying around in the mempool, anyone can observe them and trade against them.
Frontrunners (typically trading bots) can observe the unconfirmed transactions and propose their own transactions, offering a higher gas fee (transaction fee) to the miners. Miners will prioritize the frontrunners’ transactions in the proposed new block because they offer higher transaction fees. This allows frontrunners to have their transactions executed before the transactions of unsuspecting users, who had offered lower transaction fees to the miner.
This is a simplified explanation, but the point is that frontrunning on the Ethereum mempool is facilitated by how the Ethereum blockchain works right now. Miners do not have to mine transactions in the chronological order in which they were submitted. Instead, they can pick and choose the unconfirmed transactions that maximize their own rewards. That’s the financial incentive in the game theoretic architecture of Ethereum. A Chainlink blog post states that this makes blockchains ‘temporarily centralized, in the sense that a single miner gets to decide unilaterally how to order transactions in each mined block.’
The authors of the Flashboys 2.0 paper concluded that so-called Miner Extractable Value (MEV) posed ‘a realistic threat to Ethereum’ at the time of writing (2019).
Frontrunning opportunities on Ethereum increased as DeFi, which is built mainly on the Ethereum blockchain, exploded. The Ethereum mempool was called a ‘dark forest’, pointing to the arbitrage bots attempting to exploit any profitable opportunity they spot. Submitting transactions to the mempool can be ‘like a flashing “free money” sign pointing directly at this profitable opportunity’, the authors wrote.
Frontrunning may have become even more profitable since the Flashboys 2.0 paper. The Flashbots project shows almost US$250 million has been extracted through MEV/frontrunning in the past 30 days (for methodology and limitations, see here). However, the crypto community is not sitting still and is working on solutions.
Some DeFi users are frontrunning the frontrunners as a sign of their disapproval of the practice. There have been a couple of reports and explanations of the issue and how to mitigate it. And some blockchains, protocols and apps have built in design features that claim to avoid the possibility of frontrunning altogether. So when crypto investor Mark Cuban tweeted: ‘crypto has the same problem that HFTs bring to stocks, front-running is legal’, users threw a couple of crypto projects back at him that aim to mitigate or even avoid this problem altogether.
Some of these projects propose frontrunning-mitigating solutions for the Ethereum network. One example is decentralized oracle provider Chainlink (where Ari Juels, one of the co-authors of the Flashboys 2.0 paper is now Chainlink’s Chief Scientist), which discussed its Fair Sequencing Services as a solution.
The Dark Forest paper suggests that, if you know a miner, you can ask them to include your transaction in a new block and bypass the Ethereum mempool altogether to avoid the frontrunning risk by arbitrage bots. However, you need to trust the miners: miners could also frontrun transactions themselves. The FlashBots project works with a network of white-listed miners, to which users can propose transactions directly, bypassing the Ethereum mempool.
In short, there are 3 broad categories of solutions to frontrunning on-chain, according to one paper:
Why has frontrunning not been a concern on the Bitcoin blockchain?
The Bitcoin blockchain works differently than Ethereum. Bitcoin does not have the smart contract functionality that the Ethereum blockchain has. Standard Bitcoin transactions are stand-alone events, making frontrunning much less likely. The smart contracts built on Ethereum, however, can link different transactions, thereby increasing frontrunning opportunities.
Will frontrunning and MEV end with Ethereum 2.0?
No, according to research by Flashbots. There have long been plans to move Ethereum from proof-of-work consensus to proof-of-stake. In proof-of-stake, the traditional miners become redundant. If miners are no longer needed, does that mean that frontrunners can no longer ‘bribe’ miners with higher gas fees to execute their strategies? No. Someone still has control over transaction ordering and that someone can again be ‘bribed’ (financially incentivized) to order transactions in a certain way. Previously this someone was a miner; in Ethereum 2.0 it will be validators. (For a simpler description, see Coindesk.)
So, is frontrunning/MEV inevitable?
‘Not unless we choose it to be. What needs to be inevitable is that MEV becomes history because if it doesn’t, we’d better start hoping Ethereum fails or we’re all in a lot of trouble,’ said one of the early voices on frontrunning in a Coindesk opinion piece last month.